Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

This is my third question here.

For the past two days, I have asked two questions here, both of which have a bearing on my current question. I have been working on deriving an equation relating $\sigma(n^2)$ to $\sigma(n)$, if $n$ is odd.

Now, to begin with, we know that $\sigma(n^2)$ is odd, but we can't really be sure about the residue of $\sigma(n^2)$ modulo $4$, even if I know (by my current considerations) that $\sigma(n)$ is incongruent to $2$ modulo $4$. [Edit: As of today [June 22 2012], I am no longer sure about $\sigma(n) \not\equiv 2 \pmod 4$! :-(] I tried tackling the various cases that need to be considered to cover this problem in its full generality, but I think this mammoth task is too complex for me. Hence, I decided to ask this third question:

Is there a specific equation relating $\sigma(n^2)$ to $\sigma(n)$, if $n$ is odd?

Since $n \mid n^2$ and $n < n^2$, all the divisors of $n$ also divide $n^2$. Thus, intuitively, I know that $\sigma(n)$ forms part of the sum that is $\sigma(n^2)$. In other words, I have:

$$\sigma(n^2) = \sigma(n) + \sum_{\substack{d \mid n^2 \\ n < d \le n^2}}{d}.$$

(I am not so sure about that second addend, though.) Furthermore, since $I(n) < I(n^2) < (I(n))^2$, where $I(x) = \frac{\sigma(x)}{x}$ is the abundancy index of the positive integer $x$, then $n\sigma(n) < \sigma(n^2) < (\sigma(n))^2$. So then, I have:

$$\sigma(n^2) = n\sigma(n) + A, \hspace{0.10in} (**)$$


$$\sigma(n^2) = (\sigma(n))^2 - B. \hspace{0.10in} (***)$$

My obvious question at this point is: Is it possible to express $A$ or $B$ in terms of $n$, without using $(**)$ and $(***)$?

Appreciate any feedback/replies on this. Thanks!

share|cite|improve this question
There can be more terms in that sum $\sigma(n^2) = \sigma(n) + ...$. Take $n = 15$ and note that $15^2$ has $9$ as a divisor – Cocopuffs Jun 20 '12 at 22:48
Note that $\sigma(21)=\sigma(31)$ but $\sigma(21^2)\ne\sigma(31^2)$, so it is impossible for $\sigma(n^2)$ to be purely a function of $\sigma(n)$. On the other hand, allowing $\sigma(n^2)=f(n,\sigma(n))$, we should wonder how $f$ is being restricted (as $f(x,y)=\sigma(x^2)$ would be trivial): precisely what is or isn't allowed? – anon Jun 21 '12 at 1:44
up vote 2 down vote accepted

Since $\sigma$ is multiplicative we can consider powers of primes as a base case. $$ \sigma(p^a) = \sum_{j=0}^a p^j = \frac{p^{a+1}-1}{p-1} \\ \sigma(p^{2a}) = \frac{p^{2a+1}-1}{p-1}=p^a \sigma(p^a)+\frac{p^a-1}{p-1} $$ So if $n=\prod p_i^{a_i}$ then $$ \sigma(n^2) = \prod \left(p_i^{a_i} \sigma(p_i^{a_i})+\frac{p_i^{a_i}-1}{p_i-1}\right) $$ The first term in the product gives $n\sigma(n)$, but the other terms rely on the factorization in a way that I don't think will allow a simple expression in terms of $n$ alone.

share|cite|improve this answer
Thank you very much for this answer, Zander! – user11235813 Jun 21 '12 at 4:41
We can rewrite Zander's final expression for $\sigma(n^2)$ as: $$\sigma(n^2) = \prod\left({{p_i}\sigma({p_i}^{a_i}) + \sigma({p_i}^{a_i - 1})}\right).$$ – user11235813 Jun 21 '12 at 20:12

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.